-
LBNL-42345
Survey Talk-New Laser and Optical Radiation Diagnostics
W.P. Leemans Accelerator and Fusion Research Division
September 1998 Presented at the XIXhtmational Linac Conference,
Chicago, IL,
and to be published in the Proceedings
August 23-28,1998,
-
DISCLAIMER
This document was prepared as an account of work sponsored by t
h e United States Government. While this document is believed to
contain correct information, neither the United States Government
nor any agency thereof, nor The Regents of the University of
California, nor any of their employees, makes any warranty, express
or implied, or assumes any legal responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product,
or process disclosed, or represents that its use would not infringe
privately owned rights. Reference herein to any specific commercial
product, process, or service by its trade name, trademark,
manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United
States Government or any agency thereof, or The Regents of the
University of California. The views and opinions of authors
expressed herein do not necessarily state or reflect those of t h e
United States Government or any agency thereof, or The Regents of
the University of California.
Ernest Orlando Lawrence Berkeley National Laboratory is an equal
opportunity employer.
-
DISCLAIMER
Portions of this document may be illegible in electronic image
products. Images are produced from the best available original
document.
-
LBNL-42345 CBP Note-274
Survey Talk-New Laser and Optical Radiation Diagnostics*
W. P. Leemans
Center for Beam Physics Accelerator and Fusion Research
Division
Ernest Orlando Lawrence Berkeley National Laboratory Berkeley,
California 94720
Presented at XIX International Linac Conference, Chicago,
Illinois, August 23-28, 1998
This work was supported by the Director, Office of Energy
Research, Office of High Energy and Nuclear Physics, Division of
High Energy Physics, of the U.S. Department of Energy under
Contract No. DE-AC 03-76SF00098.
*
-
Survey Talk - New Laser and Optical Radiation Diagnostics” W. P.
Leemans
Center for Beam Physics, Accelerator and Fusion Research
Division Ernest Orlando Lawrence Berkeley National Laboratory,
Berkeley, CA 94720
Abstract New techniques are reported for electron beam
monitoring, that rely either on the analysis of the properties of
wiggler radiation (from static magnetic fields as well as from
lasex “undulators”, also ref- to as Thomson scattering) or on the
non-linear mixing of laser radiation with electron beam radiation.
The diffexat techniques reviewed ate capable of providing
information on femtosecond time scales and micron or even
sub-micron spatial scales. The laser undulator is also proposed as
a useful tool for non- destructive measurement of high power
electron beams. An example is given of measuring electron beam
enexgy and energy spread through spectral filtering of spontaneous
wiggler radiation [l]. A novel technique based on fluctuational
characteristics of radiation is described, for single shot,
nondestructive measurement of the electron beam bunch length [2,3].
Thomson scattering based beam monitoring techniques are discussed
,which, through analysis of the radiated beam properties, allow
non-destructive detailed measurement of transverse and longitudinal
distributions of relativistic electron beams [4I. Two new
techniques are discussed which rely on non-linear optical mixing of
laser radiation with electron bunch emission: differential optical
gating (DOG) [5] and electron bunch length measurement in a storage
ring based on sum-frequency generation [a.
1 INTRODUCTION Measurement of &he transverse and
longitudinal phase space properties of electron bunches produoed in
present and future high performance linacs [7-91, requireS
development of beam diagnostics with high spatial (micron or
sub-micron) and temporal (femtosecond) resolution. Measurement of
beam properties of high current, high power linacs [lo] requires
nondestructive diagnostics to be developed. Several diagnostics
will be discussed, which rely on direct measurement of the
properties of electron beam radiation, or on the interaction of
that electron beam radiation with a laser beam. In each of the
techniques discussed in this paper, the electron beam radiation is
generated through interaction of the electron beam with static
magnetic fields (e.g. wiggler radiation) or with electromagnetic
radiation from a lasex (Thomson scattering). Most of the techniques
can be applied more generally to other types of radiation
sources,
except when the unique property of a onetoone correlation
between observation angle and wavelength of the emission is used,
such as in radiation originating &om the interaction with
magnetic fields.
Wiggler radiation (from permanent magnets, electromagnetic
undulators and lasers) has been used for diagnostic purposes [l-3,
11-13] in a wide range of beam energies, as the radiation contains
the full signature of the electron beam. In Section 2, a technique
for measuring energy and energy spread through spectral filtering
of spontaneous emission of a wiggler will be &cussed [I] as
well as a technique for bunch length monitoring through
fluctuational interferometry of the incoherent light [2,3]. In
Section 3, experiments using radiation from laser Thomson
scattering [4] (Le. electromagnetic undulator) for beam
characterization will be reviewed. In Section 4, non-linear optical
mixing of laser radiation with radiation fro& electron beams
for longitudinal bunch profile measurements [5,6,14] will be
discussed.
2 WIGGLER RADIATION
2.1 Beam energy diagnostic The wiggler emission cone contains
information about the electron beam mean energy and energy spread
[15,16]. A series of proof of principle experiments El] have been d
e d out at the Accelerator Test Facility (ATF) at Brookhaven
National Laboratory, demonstrating wiggler- based beam diagnosis in
single shot mode, both for single micropulses and single
macropulses. The experiments were performed using a high precision
(0.08% peak amplitude rms) pulsed electromagnetic microwiggler from
MIT., with a wiggler period of 8.8 mm. The high microwiggler field
quality simplified the interpretation of the spectra defined by the
convolution over many parameters: energy spread, divergence, spot
size, matching, beam pointing and wiggler field errors. For a beam
energy of 44-48 MeV, the wiggler emission was in the visible, where
a wide range of optical diagnostics are available.
The wiggler emission profile was studied at the fundamental (532
nm). A m w (1 nm) bandwidth interference fdter was used to
spectrally filter the radiation cones, and the full transverse far
field pattern was recorded using a CCD camera. For a fixed
wavelength, determined
* Work supported by the US Department of Energy under contract
No. AC03-76SF00098.
-
by the filter, the cone radius depends on beam energy and
wiggler field strength, and the cone width contains information of
divergence and energy spread. Analytic expressions were derived,
showing that for energy spFeads realistic for the linac (0.5% FW)
at 48 MeV, divergence dominates over both energy spread and natural
linewidth at sufficiently large angles [l]. For small cones, both
effects are important. The &-field profile provides an
advantage over the spectrum in divergence sensitivity. A systematic
set of experiments was carried out to study cone response to beam
energy, energy spread, wiggler field strength, electron beam
misalignment and filter central wavelength. An electron beam
divergence of 0.25 mrad was extracted in a single shot measurement.
Examples of spontaneous emission cones are shown in Figure 1. Note
that with a wiggler length of 70 periods, sensitivity to as little
as 0.5% change in central beam energy was demonstrated.
A proof of principle experiment was carried out at the ATF, in
which the single shot spontaneous emission spectrum of the MIT
microwiggler, was studied for a range of bunch lengths (1-7 ps)
[1,3]. The microwiggler provided high brightness visible wavelength
emissions fix an electron beam energy of 44 MeV. A typical measwed
spectrum is shown in Figure 2a, revealing nearly 100% modulation
and the presence of random spikes of a characteristic width, from
which a bunch length of 2 ps was extracted. For comparison, a
simulation for a similar bunch length including the m e a s d
instrumental resolution is shown in Figure 2b. The important
features of the experimental data, the characteristic spike width
and the level of modulation, are repFoduced by the theory.
Quantitative agreement has also been obtained between bunch length
extracted from fluctuations and independent calibrations of beam
bunch length [3].
a99948MeV 1.000*48MeV
-: . .. .'. . . . . . . . . ., .t?*r.r=l.?b _ . . . . .? AYy=
0.5% . .
.. ..... . . . ..: . : . . ...* ; .-i. . . . . -:. . . . . ._ .
. . r . . . . . . ...:- . . . . i . . . .:-.-,- ;'". ......... . .
. . . . .. . . . . . . . . . . . . . - . . .* . . - . - ...:..-., *
. Figure 1. CCD images of a 1 nm portion of the far field wiggler
spontaneous emission profile, showing sensitivity to beam energy
(top pair - 48 MeV and 48*0.995) and energy spread (bottom pair -
0.5% and 1.5% FW). From Ref. 1. 2.1 Flmtuutionallnter$erometry In
1995, a fluctuational interferometry technique relying on the
incoherent contribution to the radiation was proposed [23. For a
radiation pulse to be longitudinally incoherent, the spectral
bandwidth A m must be much larger than the inverse of the pulse
duration T,, i.e. Amp >> 1. Using a bandpass filter, centered
around 0, and with spectral width 60, temporal coherence can be
imposed with an associated coherence time zmh = b-', effectively
breaking the pulse up in N independent portions where N = rp I
zcoh. From shot-to-shot, the intensity will vary on the order of l
l f i . Measurement of the variance of the intensity fluctuations
will then give a measure for N and hence T* = N I 6w.
0 200 250 300 350 400 450 500 550 600
CCD pixel (a wavelength) b!
600
c 300 .- v1
* .$ 200 C c-.l
100
0 - 0 50 loo 150 200 w ) 300 350 400
CCD pixel (a wavelength)
Figure 2. a) Single shot spontaneous emission spectrum from a
microwiggler at 632 nm, showing nearly 100% modulation of the
spectrum. Beam bunch length was extracted in a single shot
measurement from the spectral fluctuations. b) A simulation for the
same bunch length reproduces both the qualitative and quantitative
features of the data.
3 LASER SCATTERING DIAGNOSTICS
A different approach to generating radiation from particle beams
for beam monitoring is to use the interaction of the beam with high
intensity laser fields. In effect, the laser acts as an
electromagnetic undulator and the
-2-
-
properties of the emitted radiation can be accurately predicted
using an equivalent undulator model 1171. The scattered radiation
contains information on energy as well as on transverse and (for
short laser pulses) longitudinal distributions of the electron
beam.
At the Final Focus Test Beam (FFTB) at SLAC, transverse e-be,am
sizes as small as 70 nm were measured, by scanning a 50 GeV e-beam
across the intensity fringes of an optical standing wave [7]
produced by crossing two laser beams. The gamma ray yield depends
on the number of photons with which the electron beam interacts and
is therefore much larger at the peaks tban at the valleys of the
standing wave. Such resolution is beyond usual optical (e.g.
optical transition radiation or synchrotron radiation) based
methods.
A laser based beam diagnostic [4] which relies on analysis of
the properties of the scatted radiation has been developed and used
at the Beam Test Facility (BTF) [18] of the Center for Beam Physics
at Lawrence Berkeley National Laboratory (LBNL). Some of the
results of this experiment are discussed next.
To measure the transverse electron beam distribution for a given
slice of the electron beam, we scanned the laser beam transversely
across the electron beam in steps of 10 pm, by changing the tilt of
the focusing mirror. and monitored the x-ray yield on the phosphor
screen. It was found (Fig. 3) that the laser based technique and
the results from OTR were in good agreement and give a half- width
half maximum (HWHM) vertical size of 66 p. However, the
measurements for the beam edges diffet;ed and were both
non-Gaussian. From the Om data an KWHM horizontal size of 47 pm was
obtained.
l2 r OTRimaeeofebeam
3. I Orthogonal Thoinson scattering diagnostic The experiment
[4] was conducted at the BTF and used the 50 MeV (y = 98) linear
accelerator (linac) injector of the Advanced Light Source in
conjunction with a high power (40 mJ in 100 fs) short pulse laser
system operating at 800 nm wavelength. Electron bunches were
hansported using bend magnets and quadrupoles to an interaction
chamber where they were focused and scattered against the laser
beam. After the interaction chamber, a 60" bend magnet deflected
the electron beam onto a beam dump, away &om the forward
scattered x-rays. A 75 cm radius of curvature mirror was used to
focus the S-polarized ampWied laser pulses to about a 30 pm
diameter spot at the interaction point (IP) (measured by a charge
coupled device (CCD) camera at an equivalent image plane outside
the vacuum chamber).
To measure the spot size (and position) of the electron beam at
the IP, an image of the electron beam was obtained by relaying
optical transition radiation (Om) [19] ftom a foil onto a 16 bit
CCD camera or optical streak camera using a small f-number
telescope. Electron beam spot sizes as small as 35 pm rms have
beenmeas&.
During the interaction of an electron beam and lasea beam,
scattered x-ray photons are produced with energy U,, given by (for
y>> 1)
2y2hw0 i + y e u* = 2 2(1-c0sw),
where coo is the k.quency of the incident photons, w is the
interaction angle between the electron and laser beam (y=n/2 in our
experiments) , and 8 is the angle at which the radiation is
observed and assumed to satisfy 9 c
-
Figure 4. a) False color CCD image of the spatial profile of a
30 keV x-ray pulse on the phosphor screen, which is located 80 cm
from the IP; b) square- horizontal he-profile and fitting curve
(solid line), triangle -vertical line-profile and fitting curve
(dashed line) from Fig. 4 (a). The scale has been converted into
angular units.
By fitting the data (see Fig. 4) using Eq.(2), an electron beam
divergence of o,, (ow) = 6.3 k 0.2 (3.9 k 0.2) m d was found. F(K)
was adjusted to account for the spectral dependence of the x-ray
window transmission. The difference between G~~ and 0% is due to a
combination of the electron beam being focused astigmatically at
the IP, resulting in a tilted phase space ellipse (y, y’), and a
laser spot size much smaller than the vertical electron beam size.
As the laser beam crosses the focal volume of the electron beam,
the complete horizontal (direction of propagation of the laser)
phase space (x, x’) is sampled by the laser beam. However, only
electrons occupying the region in the vertical phase space defined
by the spatial oyerlap with the laser beam will contribute to the
x-ray flux. As opposed to the transition radiation based detector,
the laser beam therefore acts as an optical microprobe of a finite
region of the transverse phase space. This value of the electron
beam divergence is also consistent with an effective angular
divergence of the electron beam of 3.5 - 4 mrad obtained from
analyzing the x-ray spectra. Of course, the main difference is that
measurement of the spatial profile is a single shot technique as
opposed to measuring the x-ray spectra which requires accumulation
of thousands of shots.
Finally, since the x-ray yield is sensitive to both the
longitudinal bunch profile and the degree of transverse overlap
between the laser and electron beam, time- conelated phase space
properties of the electron beam can be studied. When an electron
bunch, which exhibits a finite timecorrelated energy spread
(chirp), is focused at the IP with a magnetic lattice which has
large chromatic aberrations, different temporal slices of the bunch
will be focused at different longitudinal locations. The transverse
overlap between e-beam and laser will therefore strongly depend on
which time slice the laser interacts with. This in turn will lead
to a time dependence of the x-ray yield varying faster than the
actual longitudinal charge distribution. To illustrate this, the
x-ray flux was measured as a function of the delay between laser
and e beam, for two different magnetic transport lattices. In both
lattices, the magnet settings were optimized to obtain a minimum
electron beam spot size in the horizontal and vertical plane (as
well as zero dispersion at
the IP), but chromatic abexrations were about 5 times larger in
the second lattice. Result of a 60 ps long scan (time step of 1 ps)
and time-resolved OTR from the streak camera for the lattice with
low and high chromatic aberrations is shown in Fig. 5(a, b).
- 1 FI) -3
5 O.* 2 i ~ o-6 0.4 .U 2 0.2 0 s o
- 1 .U
0.8
v)
% 0.6 - 0.4 -* a
h
2 0.2 s o
Time [ps]
Figure 5: x-ray yield vs. delay time between laser and electron
beam and profile of time resolved OTR image from a streak camera
for a lattice with a) small and b) large chromatic aberrations.
Whereas the temporal scan for the lattice with low chromatic
aberrations (Fig.5a) is in good agreement with the time-resolved
OTR measured with a visible streak camera, the scans taken for the
second configuration (FigSb) typically showed a 2-3 times larger
amplitude 5 ps wide peak sitting on a 20 ps wide pedestal. This is
to be compared to the time resolved OTR from the streak camera
which typically showed a 25-30 ps wide electron beam without any
sharp time structure. From lattice calculations, it is found that
an energy change on the orrler of 0.25 % would increase the
vertical spot size by a factor two at the IP, compared to best
focus, resulting in a proportional reduction in vertical overlap
between the laser and electron beam, and hence in x-ray yield.
These measments indicate the potential of the laser based Thomon
diagnostic to measure timearrelated energy changes of less than a
percent, with sub-picosecond time resolution.
It is important to note also that, due to the non- destructive
nature of the Thomson scattering technique, it might prove to be a
useful tool for the diagnosis of high current, high power electron
beams, such as for the DAHRT project [lo].
4 NON-LINEAR MIXING Another new direction being pursued for
developing beam diagnostics, is the non-linear mixing of laser
radiation with radiation from electron beams [141. Two recent
-4-
-
examples of the application of non-linear optics for bunch
length monitoring are discussed next: one in which a tightly
synchronized laser pulse is used to perform a cross-correlation
measurement and one in which the laser pulse is loosely
synchronized with respect to the elecmn beam.
4. I Luser Correlation with Synchrotron pulses Experiments at
the A&an& Light Source have recently shown [6l that a
synchronized laser pulse can be used to measure femtosecond
synchrofron pulses via fresuency up- conversion. Visible
synchrotron radiation from the ALS at 2 eV was sum-frequency mixed
in ‘a BBO crystal with 1.55 eV radiation from a short pulse (400
fs) Ti:A1,0, laser. By scanning the laser pulse in time with
respect to the electron bunch, a 16.6 ps rms bunch length was
measured, which is in good agreement with streak camera
measurements. Furthermore, the technique was shown to detect
sub-picosecond structure of the electron bunch, purposely imposed
on the bunch by co-propagating an intense short laser pulse with
the eIectron beam inside a wiggler. The laser beam, in the presence
of a wiggler field, causes an energy modulation of a slice of the
bunch via a FEL-like interaction. The energy modulation depth is
determined by the wiggler and the laser pulse strength, and the
duration of the slice is equal to the laser pulse length. By
propagating the modulated electron beam through a dispersive
section, this short slice can be separated from the main bunch,
leaving a small density &p&on in the main bunch. The
cross-correlation technique detected this few 100 fs long
depression [6].
4.2 Di~erential Optical Gating The second example relies on the
use of a loosely- synchronized laser pulse as a gate in a
non-linear medium for pulse length measurement in a technique which
is called Werential optical gating @oc) [5]. DOG uses two
non-linear media as gates and two detectors (A and B). The gate
pulse and the electron beam radiation are optically split in two
parts. The laser reaching gate B is delayedby a time S with respect
to the one reaching gate A. Under the assumption that the gate
pulse is much shorter than the radiation pulse (and an
instantaneous gate response), the signal seen by each detector can
be written
PI
(3) A(t1) OC EGIS (fl) B(tl + 6) OC EGZS(tI + 6)
where EG is the energy of the laser gate pulse and Zs(t) is the
instantaneous intensity of the radiation. From this measurement,
both the instantaneous intensity and its time derivative then
known, which allows bunch shape reconstruction. Through the loose
synchronization, the laser pulse randomly ‘‘walks” across the
bunch, much like interleaved sampling on digital oscilloscopes. In
recent experiments at Stanford University, the technique
has been applied to the pulse shape measurement of a picosecond
fi-ee electron laser source, using both an instantaneous gate and a
step function gate [51.
5 ACKNOWLEDGEMENTS I would like to thank Palma Catravas for
kindly contributing to the section on the fluctuational
interferometry and wiggler radiation measurements, R. Schoenlein
and M. Zolotorev for information on the non- linear mixing, C.
Rella and T. Smith for the DOG information, as well as my
collaborators on the laser based probing work. I also want to thank
E. Esmy for many useful conversations.
6 REFERENCES [l] P. Catravas, PhD thesis, MIT, 1998.
7132 (1995); M. S. Zolotorev and G. V. Stupakov, Roc. 1997
Particle Accelerator Conference. [3] P. Catravas et al., submitted
to Phys. Rev. Lett. [4] W.P. Leemans ‘et al., Phys. Rev. Lett. 77,
p.4182 (1996); R.W. SchoenIein et al., Science 274, .pp.236-238
(1996); W.P. Leenms et al., IEEE Journal of Quant. Elect. 33, pp.
1925-1934 (1997). [5] C.W. Rella et al., submitted to Optics Comm.
[6] R.W. Schoenlein, M. Zolotorev et al., private communication.
[7] T. Shintake, Nucl. Inst. & Methods A311, pp. 453 (1992),
Balakin, V. et al., Phys. Rev. Lett. 74, p. 2479- 82 (1995). [8] P.
Kung, H. Lihn and H. Wiedemann, Phys. Rev. Lett. 73, 967 (1994).
[9] D. Umstadter et al., Phys. Rev. Lett. 76, 2073 (1996); E.
Esarey et d., Phys. Rev. Lett. 79, 2682 (1997). [lo] H. Rutkowski
et al., these proceedings. [ll] P. Heimann et al., Rev. Sci.
Instrum., 66(2) 1885-8 (1995). [12] A. Lumpkin, Nucl. Instr. &
Meth. A, 393, 170-177 (1997). [13] E. Tarazona and P. Elleaume,
Rev. Sci. Instrum.,
[14] W.P. Leemans, P m . Advanced Accelerator Concepts Workshop,
AIP 398, Woodbury, NY (1997) [15] K.-J. Kim, A F Conf. Proc. 184,
Vol.1, 565-632 (1 989). [16] R. Barbini et al., Rivista del Nuovo C
a n t o , 13(6), 1-65, (1990). [17] K.-J. Kim, S. Chattopadhyay and
C.V. Shank, Nucl. Instr. And Meth. A 341, 351-354 (1994). [l8] W.P.
Leemans et al., Proc. 1993 EEE Part. A a . Conf., 83, ( W E , New
York, 1993). [19] L. Wartski et al., J. Appl. Phys. 46, 3644
(1975); M.J. de Loos, S.B. van der Geer and W.P. Leemans, Proc.
EPAC 94, 1679 (1994).
[2] M. S. Zoloto~v and G. V. Stupakov, SLAC-PUB-
66(2), 1974-7 (1995).
-5-